Group III nitride semiconductors are widely used for efficient blue and ultraviolet light emitting diodes, lasers, ultraviolet detectors, and field effect transistors. Due to a wide band-gap, these materials are one of the primary choices for fabricating deep ultraviolet light emitting diodes (DUV LEDs). In recent years, significant advances have been made in improving the efficiency of DUV LED. However, overall efficiencies of these devices remain low. For fabrication of DUV LEDs, it is crucial to achieve a high quality AlN buffer layer as an underlying layer for the subsequent growth of any Al-rich group III nitride semiconductor layers. However, it is extremely difficult to grow an AlN layer with high crystal quality on sapphire, silicon carbide (SiC), and silicon, which are currently the main substrates for growth of group III nitride devices.
For light emitting devices, such as light emitting diodes (LEDs) and especially deep ultraviolet LEDs (DUV LEDs), minimizing a dislocation density and a number of cracks in the semiconductor layers increases the efficiency of the device. To this extent, several approaches have sought to grow low-defect semiconductor layers on patterned substrates. These approaches typically rely on reducing stresses present in epitaxially grown semiconductor layers.
For example, one approach to reduce stress accumulation in an epitaxially grown layer relies on patterning the underlying substrate using microchannel epitaxy (MCE). Using MCE, a narrow channel is used as a nucleation center containing low defect information from the substrate. An opening in a mask acts as a microchannel, which transfers crystal information to the overgrown layer, while the mask prevents dislocations from transferring to the overgrown layer. As a result, the overgrown layer can become dislocation free. The three-dimensional structure of the MCE also provides another advantage to stress release. The residual stress can be released effectively since the overgrown layer easily deforms. In another approach, a mask is applied at a location of a large concentration of dislocation densities to block their further propagation.
Other approaches rely on epitaxially growing a group III nitride based semiconductor superlattice. The superlattice structure mitigates the strain difference between an aluminum nitride (AlN)/sapphire template and the subsequent thick AlxGa1-xN (where 0≦x≦1) layers. For devices such as DUV LEDs, thick AlGaN epitaxial layers (e.g., of the order of a few micrometers) are desirable to reduce current crowding. Using a superlattice approach, an AlN/AlGaN superlattice was grown to reduce biaxial tensile strain and a 3.0-μm-thick Al0.2Ga0.8N was achieved on sapphire without any cracks. Such a superlattice can be used to minimize the dislocation density due to varying stresses in the sublayers of the superlattice elements.
While the superlattice approaches allow some control of tensile and compressive stresses in epitaxially grown nitride semiconductor layers, the approaches do not enable epitaxial growth of nitride based semiconductor layers with uniform composition. Based on previous experience obtained from GaN growth, lateral epitaxial overgrowth (LEO) has been shown to be an efficient way for significant reduction of dislocation in GaN films. Several other technologies evolved from LEO, such as pendeo-epitaxial, cantilever epitaxy, and facet controlled LEO, have also been developed. While the above approaches work well for epitaxial growth of GaN semiconductor layers, epitaxial growth of AlN layers is more challenging due to small lateral growth of AlN films.
Another leading approach includes growth of AlN films over patterned substrates, such as, for example, patterned sapphire substrate (PSS). While the PSS based approaches generally produce an AlN layer with reduced stress and low dislocation densities, the patterning process and subsequent growth of AlN films is technologically complicated and costly.
One previously described approach proposes to reduce dislocation density between an AlGaN layer and a sapphire substrate by forming a self-organizing porous AlN layer of non-coalescing column-like islands with flat tops on the substrate. In this approach, the self-organizing porous AlN layer is grown at a relatively low V/III ratio of sixty. While the AlN layer is described as being porous, the disclosure does not define any details of such layer, nor does it disclose how the pores are connected or their sizes and densities.